Disability represents a gap between the capacities of individuals and the environments in which they attempt to complete various tasks. For older people, the ability to take care of basic personal and household needs is central to individual quality of life and has important implications for the societal costs associated with population aging. The topic of disability has grown into a significant sub-field within the general field of demography of aging, with an increase in the volume of studies and advancement in both conceptualization of and empirical evidence regarding the disability process. Studies in both developed and developing countries have identified characteristics of those most at risk of disability, assessed time trends in disability outcomes at population levels, estimated disability-free life expectancies, and investigated the nature of and risk factors associated with individual disability transitions from one period to the next. A key finding of this research is that disability is inherently dynamic, with older adults undergoing multiple transitions as they age. Yet classifications of these multiple transitions through the modeling of multi-period trajectories of disability are rare, as are examinations of how individual characteristics are associated with such trajectories. Disability trajectory research is virtually nonexistent in developing societies, even though many are aging rapidly and an understanding of health outcomes at older ages is critical, given the expected increases in formal and informal care demands that will accompany a growing number of older people. Accordingly, this project focuses on the identification of common multi-period disability trajectories among samples of elderly individuals who have been followed over time in two of the world's fastest aging populations, Mainland China and Taiwan, and investigates the associations of individual characteristics with more versus less favorable disability trajectories. A group-based modeling approach, which previously has been employed most commonly in the modeling of psychopathology and antisocial behaviors, is used to identify a small number of common disability trajectories in each population. As part of the project, the statistical methodology is being enhanced to simultaneously model disability trajectories together with the probabilities of dropping out of longitudinal surveys as a result of both mortality and other types of survey attrition. Outputs of the model are predicted disability trajectory shapes and mortality probabilities for each group, plus individual predicted probabilities of membership in each disability trajectory group. The last are subsequently modeled using multinomial logit regression and covariates representing a variety of domains, namely, demographic, socioeconomic, social support, early life, health behavior, and pathology/impairment. Finally, the project plans to conduct several other sets of exploratory analyses that advance understanding of the dynamic nature of disability, including investigation of cohort effects, time-varying covariates, and the relation of baseline pathologies to subsequent disability. PUBLIC HEALTH RELEVANCE: Disability of older adults is a major public health concern that affects both individual quality of life and societal costs of providing health care and other services. Yet studies that examine the disability experiences that individuals have over a number of years are rare in developed countries and virtually non-existent in developing countries, which are also experiencing rapid increases in their older populations. This project identifies the types of disability changes that are typical as people age in Taiwan and Mainland China, as well as characteristics of individuals who have more versus less favorable experiences.